1. SPHERES 0-G Autonomous Rendezvous and Docking Testbed Presented To DARPA Orbital Express December 2000 MIT Space Systems Laboratory David W. Miller (617) 253-3288 MIT, Cambridge MA millerd@mit.edu

2. SPHERES (AFRL-0012) CONCEPT  OBJECTIVE — Provide a testbed for long duration, micro- gravity, low risk development of metrology, autonomy and control technologies in support of autonomous rendezvous and docking for DoD and NASA missions.  DESCRIPTION — Three 0.25 meter diameter, 3.0 kilogram, self- contained satellites with on-board propulsion, processing, RF communication and metrology. — Communicates with Shuttle/ISS ThinkPads (laptops) for Ku-band (up)downlink access. — Patterned after MIT MODE (STS-40, 48, 62) and MACE (STS-67, ISS) controls laboratories. — Successfully completed prototype testing on Air Force, NASA, and MIT funded KC-135 flights in Feb and Mar 2000. — Manifested on ISS-9a in May 2002

3. Using ISS to Mature Mission Technologies  SPHERES on ISS is designed to mature algorithmic technologies (metrology, autonomy and control) for multi-vehicle autonomous rendezvous & docking.  SPHERES has access to long duration  -G that allows 6 DOF per vehicle testing under large relative motions between vehicles in close proximity.  SPHERES is a unique facility that allows algorithms at low TRL to be matured in a representative space environment — Tolerant to risk associated with low TRL since crew can replenish consumables, terminate tests exhibiting anomalous behavior, etc. — Fosters technology maturation due to crew observations, video coverage, and uplink of algorithms and downlink of data within days  R&D has gone to great lengths to simulate the space environment in the research laboratory. Now, ISS simulates the research laboratory in space.  SPHERES provides a low cost facility in space for developing & downselecting between algorithms for OE

4. Current Testing Using SPHERES  Single SPHERE maneuver control on the KC-135 in February 2000  Multi-SPHERE formation flight coordination on the KC-135  Multi-SPHERE rendezvous and docking in the SSL 1-G laboratory  Future upgrades — Emulate docking with a target vehicle in free drift — Emulate a thruster failure in resupply vehicle — Once docked, autonomously identify new inertia properties and reconfigure control — Replace velcro with more advanced docking capability

5. Current Testing Using SPHERES  Single and Multiple SPHERES units maneuvers in the KC-135, February and March 2000 — Testbed Validation — Initial Formation Flight

6. Current Testing Using SPHERES  One-g SSL Laboratory Experiment — Development of 3DOF rendezvous and docking using global coordinates

7. Relevance to DARPA’s Orbital Express (I)  Orbital Express must demonstrate three key features — (1) fuel transfer, (2) avionics upgrade & (3) routine auto. rendezvous & docking — These are essential to replenishment, inspection, and repair of existing assets to lengthen life, recover from partial failures, upgrade technologies, and identify causes  Fuel transfer demonstrated in Shuttle’s payload bay  Avionics upgrade performed by astronauts on the Hubble Space Telescope: human-in-the-loop  Rendezvous and docking demonstrated in limited forms — Manual human-in-the-loop with Shuttle to MIR and ISS — Automated with human-supervisory- control of Progress to MIR  Orbital Express requires routine autonomous rendezvous & docking — Without human supervision — With ability to adapt to low level anomalies — That can accommodate cooperative, non- cooperative, and eventually un- cooperative target vehicles  Routine autonomous rendezvous & docking is the most immature

8.  Routine autonomous rendezvous & docking raises several questions — How does the problem change as different information becomes available from the two vehicles? — Both vehicles communicate and coordinate their motion — Target nulls residual velocities while docking vehicle performs all maneuvers — Docking vehicle must match residual motion of non- cooperative target — Can safe mode and recovery logic be developed that requires minimal human intervention? — Can autonomous close proximity operations avoid collision and plume impingement?  These define a wide design space which must be explored before committing these algorithms to OE flight demonstration  The SPHERES Autonomous Rendezvous and Docking Testbed can be used to mature these algorithms in an environment that: — Provides long duration micro-G for close proximity operations — Is risk tolerant by allowing IFM and replenishment of consumables — Has access to video coverage and Ku-Band (up)downlink facilitating iterative algorithm refinement — Has low cost and high visibility Relevance to DARPA’s Orbital Express (II)

9. SPHERES (AFRL-0012) DETAILED OVERVIEW  FLIGHT SYSTEM — Flight H/W (fits in 1-1.5 middeck lockers) — 3 SPHERES, 4 metrology transmitters, 1 laptop (GFE) — SPHERE satellite contents — CO 2 propulsion tank, RF communication, IR-ultrasonic global metrology, Inertial Measurement Unit (IMU), AA battery power — Researcher uplinks algorithms, crew down- loads from laptop, crew initiates test and replenishes consumables, crew downloads and downlinks data to ground, researcher reviews data and refines algorithms, researcher uplinks refined algorithms. Cycle completed in days.  STATUS — Currently manifested on ISS-9a in May 2002 for 4-6 months on ISS. — High fidelity prototype built & operating in lab & KC-135, Phase 0/1 Safety Package complete, EMI tests conducted  PRIORITY — DoD SERB rank 15/34 — AF SERB rank 9/14  FUNDING NEEDED — Need $900k to transition from high fidelity prototype to operation on ISS — Flight hardware fabrication, STS-ISS integration, operations — Potential non-DARPA sources include NASA ST-6 proposal & SBIR, and Lockheed & AFRL

10. SPHERES Team Capability  MIT Space Systems Laboratory — David W. Miller — Formation flight, rendezvous and docking research in support of Techsat21, ST-3, Terrestrial Planet Finder — Design and PI of 0-g dynamics and controls laboratories MODE STS-40, 48, 62 DLS on MIR MACE STS-67, 106, ISS — Jonathan P. How — Formation flight, differential GPS, robust control — Brian Williams — Spacecraft autonomy, remote agent, Livingstone autonomous model-based diagnosis on DS-1  Payload Systems Incorporated — Developer and integrator of experiments in human-rated space platforms (Shuttle, MIR, ISS)  The fact that our facilities have more reflights first flights is testimony to the versatility of, and demand for, our dynamics and controls laboratories